The semiconductor technology has advanced centering around silicon. The scale of integration has increased from transistor devices to IC (integrated circuit) and VLSI (very large scale integrated circuit), and the increase in scale of integration is expected to continue in the future. With an increase in scale of integration, however, it is feared that corresponding increases in the operation speed is limited by the delay of electric signal through wiring. As the counter-measure for this, optical interconnections are drawing attention, and the monolithic integration of silicon based electronic devices with III-V compound semiconductor optical devices is considered the important basic technique whereby to realize optical interconnections.
For use in forming the III-V compound semiconductor optical devices on a Si substrate, the following two methods are being studied. One is the so-called super-heteroepitaxial method for epitaxially growing a III-V compound semiconductor, such as a GaAs or InP, which differs in lattice constant from Si, on a Si substrate, and forming an optical semiconductor of an AlGaAs, InGaAs or other system formed on top of the grown III-V compound semiconductor. An example of this method is disclosed in “Gallium arsenide and other compound semiconductors on silicon” by S. F. Fang, K. Adomi, S. Iyer, H. Morkoc, and H. Zabel in J. Appl. Phys. 68(7), Oct. 1, 1990.
The other is a direct bonding method in which optical semiconductors are first grown on a GaAs or InP substrate and then bonded to a Si substrate, and then the GaAs or InP substrate is removed. Such direct bonding method is disclosed in “Semiconductor lasers on Si substrate using the technology of bonding by atomic rearrangement” by Y. H. Lo et al., Appl. Phys. lett. Vol. 62, pp 1038-1040, 1993.
Regarding the materials used to date for the III-V compound semiconductors, as written in H. C. Casey, Jr. and M. B. Panish, “Heterostructure Lasers—Part B”, Academic Press, New York, 1978, pp. 8-9, binary compound semiconductors made up of Al, Ga or In as a group III element and P, As or Sb as a group V element, and alloy semiconductors including those elements have long been used. With progress of the crystal growth techniques, N-containing alloy semiconductors have recently come to be formed, such as GaNP (J. N. Baillargeon, K. Y. Cheng, G. E. Holfer, P. J. Pearch, and K. C. Hsieh, Appl. Phys. Lett. Vol. 60 pp. 2540-2542, 1992) and GaNAs (M. Weyers, M. Sato and H. Ando, Jpn. J. Appl. Phys. Vol. 31, 1992, pp. L853-L855). This has widened the range of choice of materials. In addition, a case where the N-containing alloy semiconductors are grown epitaxially on Si substrates is disclosed in JP-A-1-211912. When an N-containing alloy semiconductor is actually applied to form a semiconductor device like a laser diode, it is necessary to design a multi-layer structure by calculating the band gap and the amount of lattice strain. Because the N-containing alloy semiconductors show a huge bowing occurring in the band gap owing to N's extremely high electro-negativity, which will be described later, such special consideration which is not needed for the conventional alloy semiconductors is required in designing the band gap of the multi-layer structure of the N-containing alloy semiconductors. However, in all the cases of the N-containing alloy semiconductors grown to date, only an epitaxial monolayer is grown on the substrate crystal and, therefore, there has never been a case where N-containing alloy semiconductors are deposited in a multi-structure and applied to form semiconductor devices.
Both in the super-heteroepitaxial method and the direct bonding method, the lattice constant of the materials constituting the optical semiconductor device differs greatly (more than 4%) from that of the Si substrate, so that there is a problem that a misfit dislocation occurs in the crystal near the interface between the Si substrate and the III-V compound semiconductor. Another problem is that due to the difference in thermal expansion coefficient between the Si substrate and the III-V compound semiconductor, the dislocations, which have been generated in the cooling process after a heating process in the epitaxial growth or the bonding, move and increase. As a result, there are problems regarding the characteristics and the device lifetime of the optical semiconductor device produced. Therefore, the monolithic integration of silicon based electronic devices with the III-V compound semiconductor optical devices has not been put into practical application.